Static and Dynamic Scheduling Method of Demand-Responsive Feeder Transit for High-Speed Railway Hub Area

Demand-responsive feeder transit (DRFT) is an emerging urban public transport mode with the advantage of offering flexible door-to-door services in the high-speed railway hub area. However, the existing bus scheduling schemes can hardly meet personalized and diversified passenger transfer demands in the station-city integrated high-speed railway hub, reducing the attractiveness of DRFT. This paper studies the DRFT scheduling problem considering static and dynamic travel demands under the background of mobility as a service (MaaS). An information-based DRFT system framework is proposed, where the K-means clustering algorithm is implemented to select target bus stops from regional road networks for passengers to get on and off. A two-stage mixed integer programming model is first formulated to generate operational routes and optimize the static and dynamic scheduling before and after departure. The objective functions reflect the operating benefits of public transport enterprises and the travel costs of passengers, and the demand characteristics in different driving directions are taken into account in the model. Then, an improved genetic algorithm is developed to solve the model, which is called the genetic algorithm-exact algorithm (GA-EA) in this paper. Finally, the proposed model and algorithm are evaluated using the case study of the Nanjingnan Railway Station area. The experiment results show that the optimal scheme can provide a 100% demand-response rate, reasonable service time, and valid driving routes. In addition, compared with GA, the average search time of GA-EA is shortened by 43.5% and the total objective function value is increased by 2.16%. The findings in this paper can provide practical guidance on DRFT scheduling and improve the efficiency of bus feeder service.

Automated summary

This paper studies the DRFT scheduling problem considering static and dynamic travel demands under the background of mobility as a service (MaaS). An information-based DRFT system framework is proposed, and an improved genetic algorithm is developed to solve the model. The experiment results show that the optimal scheme can provide a 100% demand-response rate, reasonable service time, and valid driving routes.